Lockdown Mobility Network Analysis

Final project for the course of Social Network Analysis at the University of Trieste, A.A. 2019-2020 held by Prof. Susanna Zaccarin and Prof. Domenico De Stefano.

Analysis of shifts in mobility patterns during Italy’s COVID-19 lockdown. Built by Gabriele Sarti and Enrico Fallacara.

See data/README.md for specifications about the data used.

Visualizing Mobility Shifts

Mobility in Italy during the 2020 COVID-19 Lockdown

Simulating Mobility Shifts Progression in the Future

Simulation of Italy's mobility after the COVID-19 pandemic

Guidelines

Taken from Moodle:

The exam will consist in the presentation and discussion, in groups of 2 up to 4 students, on the analysis (description and model fitting) of a real network dataset, explaining the working steps and the obtained results. The writing of a short report is also requested, with the commented R code.

During the presentations, few questions will be asked to assess the individual contributions and preparation on the topics of the course.

Students are asked to carry out the following tasks:

  1. Perform a descriptive analysis of the assigned network(s) (and attributes) dataset using the indices and tools learnt in the course.

  2. Perform an exploratory analysis by finding communities (or alternatively blocks by means of blockmodeling).

  3. Specification and estimation of an ERG model on binary network, choosing an appropriate threshold to dychotomise the weighted relations. Since networks are large-sized, please use a meaningful sub-network to fit the ERGM in a reasonable amount of time.

  4. Perform analysis for each network and compare results

Delivery deadline of the short report (max 10 pages) and the commented R code: 29th June, 11.59 p.m.

TO-DOs

1

  • [x] Review techniques of descriptive analysis presented during the course
  • [x] Select a metric to weight edges among those available
  • [x] Perform descriptive analysis of the 3 networks in an R script
  • [x] Write report section comparing similarities and differences in the analysis
  • [ ] Add comments to the script

2

  • [x] Review techniques of exploratory analysis for community detection
  • [x] Perform community detection for the 3 networks in an R script
  • [ ] Write report section highlighting relevant communities and techniques used
  • [ ] Add comments to the script

3

  • [ ] Review exponential random graphs presented during the course
  • [x] Select subnetworks for all three dataset to fit ERGs (Giant component, non-looping edges)
  • [x] Select additional features to improve the quality of fit (Mortality, Tot. population and Tot. positive per province)
  • [ ] Perform the modeling R script
  • [ ] Write report section showing the structure and performances of various models
  • [ ] Add comments to the script

Useful sources:

  • Italian Public Datasets (especially Automobile Club Italia about available number of vehicles)

  • OECD Stats (see Regions and Cities > Regional Statistics, ITA provices are “Small regions” TL3)

4

  • [ ] Perform additional analysis to gain more insights about the mobility changes (find something related to the analysis a geographical networks?)
  • [ ] Write final report section

Useful sources: